Themis AI: Improving AI Reliability

Themis AI: A Leap Forward in AI Reliability

Themis AI, an MIT spinout, is pioneering advancements in AI reliability by addressing model uncertainty and enhancing performance in critical sectors. The company's innovative Capsa platform works with machine-learning models to identify and correct unreliable outputs, ensuring AI systems function correctly even in high-stakes industries.

Founded in 2021, Themis AI has already made significant strides, assisting telecom companies with network planning and automation, and helping oil and gas companies analyze seismic imagery using AI. The company's research on reliable chatbots further underscores its commitment to improving AI systems.

The Capsa Platform: Enhancing AI Accuracy

The Capsa platform, developed by Themis AI, is designed to quantify model uncertainty and rectify outputs. According to co-founder Daniela Rus, Capsa identifies a model's uncertainties and failure modes, enabling it to enhance performance and ensure correct functioning.

Alexander Amini, another co-founder, emphasizes the importance of enabling AI in high-stakes industries. "AI errors in critical sectors can have severe consequences," Amini notes. "Themis AI aims to enable AI systems to predict their own failures, thereby improving reliability."

Research and Development

Themis AI's research and development efforts are deeply rooted in years of studying model uncertainty. In 2018, Rus's lab received funding to study the reliability of machine learning for autonomous driving. The team also created an algorithm to detect and eliminate racial and gender bias in facial recognition systems by reweighting training data.

In 2021, the co-founders demonstrated that a similar approach could help pharmaceutical companies use AI to predict drug candidate properties. This breakthrough led to the founding of Themis AI, which continues to collaborate with companies across various industries, particularly those using large language models.

Future Impact and Applications

Themis AI is currently exploring Capsa’s ability to improve accuracy in chain-of-thought reasoning, where large language models explain their reasoning steps. Stewart Jamieson, Themis AI's head of technology, suggests that Capsa could guide reasoning processes to identify the highest-confidence chains, potentially improving the overall AI experience and reducing computation needs.

The company is also in talks with semiconductor firms to create AI solutions that function outside cloud environments, offering efficient edge computing without sacrificing quality. This approach allows edge devices to handle most tasks, forwarding uncertain outputs to a central server.

Pharmaceutical companies can also leverage Capsa to refine AI models for identifying drug candidates and predicting clinical trial performance. Amini notes that Capsa can offer insights into whether predictions are supported by training data, potentially accelerating the identification of the strongest predictions.

Conclusion

Themis AI represents a significant step forward in AI reliability, addressing both the potential and concerns of AI technology. With its innovative Capsa platform and ongoing research, Themis AI is poised to make a real-world impact, ensuring that AI systems function accurately and reliably in critical sectors.